Sensor Fault Diagnosis in Autonomous Ships

Autonomous ships heavily depend on their sensor systems for safe and efficient operation. When these critical sensor systems are compromised by faults, the entire autonomous operation is put at risk. Detecting and accurately estimating the magnitude of such faults becomes imperative to ensure the reliability and safety of autonomous ships. In response to this challenge, this paper presents a robust methodology built upon adaptive Kalman filter with forgetting factor to estimate the magnitude of sensor faults. What sets our approach apart is the innovative perspective taken towards fault diagnosis. Instead of treating the fault as an additional state variable within the system, we directly estimate the fault magnitude based on the available measurements. Our approach is demonstrated through extensive simulations, showcasing the effectiveness and resilience of the proposed method. The results highlight its potential to significantly enhance the dependability of autonomous ships in the face of sensor faults, contributing to their continued success in a wide range of real-world applications.